EARLY DETECTION OF SIMILAR FAKE ACCOUNTS ON TWITTER USING THE RANDOM FOREST ALGORITHM

A major issue for social-media platforms is the problem of fake accounts with different aims and objectives. A similar fake account is like having access to someone's specific identity, which may impact that person's life in the real world. Artificial intelligence is leading in dealing with these issues, because its machine-learning methodology can provide early detection of similar fake accounts on Twitter. In this work, we analyze the early detection of similar fake accounts on Twitter using the Twitter API Application Programming Interface mainly uses the following features based the confusion matrix: default_profile, default_profile_image, friends_count, statuses_count, followers_count, listed_count, listed_count, profile_background_image, verified, name, and id. These are the data features we chose to enable early detection of similar fake accounts, and we used the random forest algorithm in the model. We find that overall the model works better than other approaches, and the random forest algorithm provides impressive results even in the validation phase. The random forest results depend upon the features selected to identify the similar fake accounts. The model produced impressive results in the early detection of similar fake accounts on Twitter